- Email: [email protected]
Digoxin reduces cardiac sympathetic activity in severe congestive heart failure
JACC Vat. 28, No. 1 July 19!36255-61
155 HEART FAILURE
Digoxin Reduces Cardiac Sympathetic Actitity in Severe Congestive Heart Failure GARY E. NEWTON, MD, JEFFREY H. TONG, PHD,* ANNE M. SCHOFIELD, BScN, ANDREW D. BAINES, MD, PHD,* JOHN S. FLORAS, MD, DPHI~ FACC, JOHN D. PARKER, Toronto,
Ontario,
MD
Canaak
UAm call l2vllid 1996;2&155-61)
The syndrome of congestive heart failure is characterized by generalized and organ-specilk increases in efferent sympathetic activity, with the greatest increase in sympathetic out&w being directed to the heart (1). Increased sympathetic out&w to the heart may have several adverse effects in&xiii downregclation and desensitization of cxdiac beta-adrenergic receptors (2), impaired function and viability of cardiac myocytes (3) wd a predisposition to ventricular arrhythmias (4). Recent evidence (5) demonstrates that increased cardiac adrenergic drive is stron& predictive of mortality, at least in patients with heart failure undergoing evaluation for heart transplantation. Despite the adverse consequences of sympathetic activation
directed to the heart, the modulation of cardiac sympathetic outflow by medical therapy in the setting of heart failure remains largely unexplored. An agent that may favorably reduce central sympathetic outflow to the heart is digoxin. Cardiac glyadeq by increasing arterial baroreceptor and cardiac receptor nerve d&barge (6-ll), cause an &ease in efferent parasympathetic activity and a k&ction in central sympathetic outlbv (12). In animal models glycosides reduce cardiac effewnt sympathc’k activity (13,14) and increas cardii catecholamitse stores (15). In patients with heart failure digoxin reduces plasm norepkphrine concentration (1619), an approximation of gene&ized sympathetic tiivity. &diacg@&deshavealsokenshown(M)toreducesympatheticmxvetra&toskektalmusde,adkectmeasweof TyTtp&etic~tothe~.The’eilectofdigrednm !$ympdeticoutfknvloth!ehearlinhtmlanswithcoagestiwheart faihmremaiE3lmlmom. Predii which patients will have increased cardiac gympathetic o&low is dilkdt bemuse the afferent signals leading totbbdisaubanceate~~~defined(21).Lthesen~0f heart fake, efferent sympathetic activhy to skeletal musck is iSJUtWSiSlpulmonaryarterypressUrettSSd diredyrdatedto left vennicular lwlg pessures (2523). sid@. sympatbetk oll&vwtothehearias~bycafdiacnorepioephriae spjilaver,isalso p”ssures (2% 07354@97c96m5M m sa735-la97t%)fJtups
JAC’C VUI. 28. No. I July IYWxl.%-61 certain, cardiac kympalt.elic activalion in congeslivc ilcart failure is associated with elevated central filling prrssu,t .:. We hypothesized that digoxin would reduce cardiac sympathetic activity in patients with heart failure and underlying sympathetic activation. To test this hypothesis we studied a group of patients with heart failure and elevated filling pressures, because these patients have been reported (21-23) to have increased sympathetic activity. For comparison, we also, studied patients with heart failure and normal filling pressures and patients with normal ventricular function. To assess the effect of digoxin on sympathetic outflow IO the heart. we utilized the steady state radiotracer technique (24) to measure :he cardiac norcpinephrine spillover response to a single imr:!venous dose of digoxin.
Methods Patients. Twenty-four patients with heart failure (22 men, 2 women; mean age 2: SEM 58 f 2 ycan) with a left ventricular ejection fruction ~35% (radionuclide angiography [n = lb], cchocardiography [n = I]) participated in this study. Four patients were in New York &art Association functional class II and 21) were in class 111 c: IV. Twelve patients with elcvatcd left ventricular end-diastolic pressure (>14 mm Hg) were designated as Grrncp A. The etiolog of heart failure 1~1this group was ischrnric in nine and idiopathic in three patients. TrGIImenl included diuretic drugs (II = Id), angiotensclconverting enqme inhibitors (n 7 IO), digoxin (n = 9) and amiodarorre (n := 2). Five palicnts with normal filling pressuro (left ventricular end-diastolic pressure 514 mm Hg, Group R) also received digoxin. The etiology of heart failure in this group was ischl:mic in thlce and itliclpathir in two patients. ‘frcatmcrtf included diuretic drugs (n L ?). angiotcnsinconvelting enzyme inbihitors (n 2 2) ilnd digoxin (I: = 2). Scvcn patients with hem!Klynamic dat:B similar to those of Group B (Gnru co~rr~!gro~p) underwent an identical protocol without the administration of digoxin. The etiology of heart failure in this group was ischemic in three and idiopathic in four patients. Treatment included diuretic drugs (n .= 7). angiotensinconvcrting cnaymc inhihiton (n 2 h), digoxin (II .7) and amitdurcmc (n 2). Paticnls participating in thr dig&n protocol who were rccciving maintenance digox& thcrapj had this medic&ion withheld for a minimrlm of 7 days before the sludy date. The nor& r’c~nr~~~~br~~nc~/ic. 1grvrrp included five men and ow woman (mu. I LI~C 50 ! 5 yciln). Three paticms had ct)t~on;l~ artcry JIwa’iL’. bul M)IIC hild a hishfry of myncardial mfarstion. Medicid thrrapy included hta-adrcncrgic blocking agents (II = 2). calcium channel hlockmg agents (n = 2). and an angiotensinccmvertting enzyme inhihitor (n .= !), IX protocol was approved by rhe phical Review Committed for Human Ewiin;intation of Ihc :‘nivcrsicj of TG romo. Writiea informed conwnt y/a\, trhtaincd in ,111cuses“~. Shady pr&cai. Diagntatic right and fcft hc:lri *.atheterilation, without udation. was performed fnr.1 the femoral ap proaeh. A 7F ?lumnlil~~~~m caarlctcr (rvp‘ (‘G-?tJ-Wf3.
Wchstcr Ishoratories) was then placed in the coronary sinus from an antecubital vein for dew measurements and sampling. A 7F micromanometer-tipped catheter (Milbu Industries) was placed in the left ventricle for assessment of contractility (peak positive first derivative of left ventricular pressure (dP/dt) and filling pressures. The patient was then left undisturbed for a minimum of 20 min and until an intravenous infusion of tritium-labeled norepinephrine reached steady state concentration in plasma. Al that point, hemodynamic variables and total body and cardiac norepinephrine spillover rates were assessed. Digoxin (0.25 mg in IO ml 01 5% dextrose in water), was then administered intravenously over 6 min with use of a Harvard pump. Hemodynamic and norepill. ihrine spillover mcasuremcnts were repeated 30 mm later. In the time control group, the same hemodynamic and spillover i;*sasurements were performed at baseline and repeated at 30 min. Patients were classified by left ventricular enddiastolic pressure measurements into either Group A (>I4 mm Hg) or Group B (-14 mm Hg) on the basis of hemodynamic data recorded before administration of digoxin. This catego:ization was performed before the biochemical analysis. Norepinephrine spillover merwwemeots. Sympathetic activity was estimated according to the radiotracer technique of Esier et al. (24). This method provides the rate at which norepincphrine appears in plasma and is therefore an indirect measure of effcrcnt sympathetic nervous activity. To assess norepincphrine spillover, a ...~~dy state plasma concentration of :I tracer dose of tritium-labeled norepincphrine is required. Total hrldy norepincphrine spillovcr is cala:latcd as follows (L24): Tot:11 h)dy NE ‘:pillwer (pmol/min) [‘IlINE infusion rate (dpm/min) = Plasma NE &cific activity (dpmipmol) where NE = norcpinephrine, [JH]NE epincphrine, and dpm = disintegrations labclrd norcpincphrine. C’ardiac norcpincphrinl: spilltwcr is the Fick principle, with a correction tritium-labeled norepinephrine across
’
= tritiilm-labeled nor. per minute of tritiumcalculated according for the extraction the heart (1,24,25):
IO of
whcrc Nt.,.,, - tramcardiac fractional extraction of tritiumlab&d norrpinephrinc. (JH]NE,,,, and t’HINE,, = arterial and coronary sinus tritium-labeled norepinephrine conccnrralions. respectively. NE, and NE,, = coronary sinus and -rtcr~al norcpincphrinc concentr&ons. respcctivcly, ,UI~ I’SPI: = coronary sinus p&ma How calcufatcd from the hr‘nrattrrit and coronary sinus blood Row. A nwc dctaikd cbaluation t$ cardiac mwspinephriw kineticswus pwf&xrnwd in Md:hc I2 patients in Gruup A sld
JACC Vol. 24 No. 1 July 1996:155;61
in all 7 in the time control group. In these patients, the cardiac spillover of the intraneuronal norepinephrine metabnlitc, dihydroxyphenylglycol (DHPG), was calculated as the product of its venoarteriaf gradient and coronary sinus blood flow (25.26). Cardiac spillover of tritium-labeled DHPG was also calculated by the same method (25.26). Tritium-labeled DHPG spillover, which can only occur aa a result of neuronal uptake of tritium-labeled nore,inephrim, provides an index of neuronal re-uptake of norepi’ ephrine (25,26). Radiotmcer pw~watioo and ioliis!sn. Trifium-iabekd norepinephrine (L-I?J,~‘H]NE, New England Nuclear) was diluted in sterile 0.2~moffliter acetic acid in saline solution with IO-mmoi/liter ascorbate, sterilized by filtration through a 0.22~pm Millex-GV filter and packaged in 4-ml aliquots that were stored at -70°C until use. The radiotracer was administered intravenously as a IQ@Ji priming bolus followed by a continuous infusion of 1 Lo 1.2 PCiimin. Control mcdsurements were performed ~60 min after initiation of the radiotracer infusion to ensure steady state concentration in plasma (27). Analysis of plasma cateehulamines. Samples were transferred immediately to ice-chilled tubes containing an anticoagulant (ethylenediaminetetraacetic acid). The plasma was separated by centrifugation at 4°C and then stored at -70°C until assayed. After the addition of dihydroxybenzlamine (DHBA) as an internal standard, the catechols were extracted by adsorption on alumina. After elution from the alumina by the addition of O.Zmol/liter perchloric acid, the cacechols were separated by isocratic ion pair chromatography on a reversed phase column. The high performance liquid chromatography system consisted of a multisolvent delivery system model 6OOE pump, a model 7171 automatic sample injector (Waters Associates) and a Zorbax %-Cl8 4.6 X 1%mm column (Chromatographic Specialties Inc., Brockville, Ontario. Canada). Detection of the compounds was performed by a model Coulochem II
DIOOXIN
IN CONGESTIVE
Table I. Baseline Characteristics
NbWiDN ET AL HEART FAILURE
157
of the 30 Patients Patients With Heart Failure
Patients With Normal Ventricular Fmmion (n = 6) .-.-.LVEF sv c 5 HR 6v.!3 14: 2 PAmcm Cl 2.7 t. 0.1 L dilJP II r2 153 2 10 F-h FA PP 7725 i377tw tdP/dt CANFSP III3 f 41
Group A (n = 12) ?I + 1. tl6.tJ 30 t 3’ l.x+u.I* 25 t 2’ Es+7 56 t 4’ vwtvw 263 ? 70
GwB (n = 5) 31 + ,(4 : II 2 2.7 + Vtl 137: 76 + l020? 151 t
3’ 4 2t 0.3t 7 4 122 5s
Time ContFol Gmup (n = 7) 152 I’ sn? IO 33 + 5’ 1.8 ? 0.2 24 t 4” 1t1+7 5427 706 ‘i 54’ 33v ? 73
‘p < 0.0s vcmu nwmal gr”“uup.tp . UM YenIl\ alup A. Left ventrictdw ettd-dimtalic pwsure (LS’EDP [mm Hgl) was not annpted beween Group A pqd Croup 6. Data presented are mean value t SEM. CANESP = cardiac norepmephrine spillover @moUrnin); Cl F cardiac index (litem/min per m*): FA PP = femural artery pulse pressure (mm Hgj: F& = femoral artmy systobc prwure (mm Hg); HR = bean rate (beztsimin); LVEF = left vetttktdat ejection fraction (a); P&.. = mean pulmonary attery pressure (mm fig); tdP/dt = peak positive lint derivative of left ventricular pressure (mm Hgk).
tion) were recorded on a strip chart recorder. Measurements were obtained by averaging 2 I5 cardiac cycles. The coronary sinus thermodilution flow catheter was positioned under lluoroswpic guidance and correct placement was confirmed with radiographic cOntrast injection. The external thermistor was advanced 22 cm past the ostium of the coronary sinus to avoid reflux of right atriil blood (28). A amstant position was maintained during the study by comparison with anatomic landmarks. Room temperature 5% dextrose in water was infused at 35 ml/min by Harvard pump for coronary blood flow measurements, which were performed in triplicate according to the method of Ganz et al. (29). Statlstkal natysi~. All data are presented as mean value -t SEM. Statistical analysis was performed in SAS (SAS release 6.10, SAS Institute Inc.). Between-group comparisons of baseline characteristics were ma& by a one-way analysis of variance adjusting for multiple comparisons (Bonferroni). Within-group and between-group comparisons of changes between control measurements and 30 min after digoxin administration (and time control) measurements were made using a multivariate anatysii of variamc (MANOVA) with appropriate contrast statements for all mmparisons. A p value of less than O.O!i was required for statistical signitkance.
l
Results
Baselii characteristics (Table I). The 12 patients with heart failure in Group A (left ventricular end-diastolic pressure ; I4 nn Hg) had a mean left ventricular enddiastolic pressure of 25 C 2 mm Hg (rzngc I6 to 3x). The tive patiena in, Grarpr3(cfiwXtricu!ai&-diast&peswe~I4mmHg)had ameankftventrkularer&%astolicpressumof9~ ImmHg (range4tol2).InamparisonwitbGroupB.thepatientsin
158
NEWTON ET Al.. DlODXlN IN CONGESTIVE
JACC Vol. 28, No. I Jul) 1%155-61
HEART FAILURti
Table 2, Hemodynamic, Plasma (Btecholamine b---e-
and Norepinephrine
Spitlower Responses
Pv~irnlr Wilh Normal Vcntriculur Fun&n (n b) ---c Dig
Gnq A (n 12) ------
--
Pntirntr With llcarr Fadurc
-~ c
Dig
---_-----c.
:inmp B In -. S)
Time Ctihlml Group (w= 7) Dig
c
30 min
HR 69 -r 3 69 % 4 x6+5 MSi 5 8454 Xl 24 98 2 10 % 2 IO LVEDP II +-2 102 1 2522 25-3 9+1 72 I 24 2 4 23 5 4 h‘% 153 _c IO 155 + 8 12827 132 5 6 137 + 7 141 2 IO 11827 118~h FA PP 77 -r 5 80%6 56 + 4 59 _+4 7b “- 4 79i-9 54 c 7 56 f 7 +dPidl 3.377 +_99 1381 2 92 99Glc9u 1027 + 55 1020 + 121 1053 f 131 7Ob t 54 716 c 58 CSBF I43 f II 14x L 15 176 f !I IbS .Z ?I 17s t I4 Ihl + 14 185 2 35 IYI ?z34 NL,, 1.1 :o..l I,/ ? 0.2 2.2 5 0.4 I.0 i 0.3 I .4 ! 0.4 I .4 t (I.4 3.2 ? 0.X 3.0 c 0.x W I.4 t 0.4 1.4 t Il.3 3b? 1.1 3.2 r 1.w 2.0 c (I.7 2.5 f 0.W 4.5 ? o.ii t.b ?: 0.8 TBNESP 3.23 t Oh9 3.34 t 0.5v 4.02 ‘f 0.68 4 44 f Rb5 4Ih? 1.m 4.62 * I.32 b.25 ? 0.57 6.44 z 0.67 NE,., (I.71 i: (1.07 0.7l1 t o.oh 031 1 0.03 0.61 ? rl.tnt o.sv * 0.05 03x 3 0.07 0.54 f: 0.07 0.50 + 0.07 CANECL 71 t I4 7H t II ho t x 61 + 4 bV f 7 61 *_ 7 56 i- 5 5b+ 5 CANESP 103 .r 41 vu c 37 2h3 ? 70 2IX + h3 I51 + 55 IV1 t 6X' 339 c 73 339 2 65 -..-*p < OJI5 and tp < (I,01 wrsus clmtwl value for xithiwgroup compwium,. Data presented we mean value + SEM. C = conlml; CANECL = cardiac norepincphrinc clearance (ml/mm): CSBF = coronary 6iw.s hltd Row (mllmin); IJig = digoxin; NE,, and Nf&,. = arterial and coronary rinus norepinephrine waeeatration (nmcl4iter. respeclivrly): NE,,, = tramcardiac vu dm-labeled norcpinephrine cxtrdclion ($6): TBNESP = lotal body norepinephrinc qpillover (nmol/min): other ahhrtvtatmn~ 4s m Talk 1.
Group A had significantly lower values for cardiac index and higher mean pulmonary artery pressures. Although the mean value for cardiac norepinephrine spillover was higher in Group A than in Group B, the difference was not significant. Baseline hemodynamic and nemochemical values in the time control group with heart failure were similar to those in Group A. Baseline characteristics for all groups are summarized in Table 1. Humudynamic rcspunses (Tabie 2). There were no changes in any hemodynamic variable after administration of digoxin in Group A, Group Band the normal ventricular function group. There were also no hcmodynamic changes over 30 min in the titnc control group. Crrdinc and tutal body norepinephrlne spillover responses (Table 2. Fig. I and 2). In G”tuup A, there were no significant changes in arterial uorepinephrine concentration, total body norepinephrine spillover or total body norepinephrine clearance after administration of digoxin. However, coronary sinus norepinephrine concentration decreased from 3.6 t 1.1 to 3.2 4 1.0 nmol/iiter after digoxin (p < 0.01). Similarly. cardiac norepinephrine spillover dccreascd from 263 i’ 70 to 21X L 62 pmol/min after digoxin (p e.: Wttl). The indivtduat cardiac norcpincphrine spillover rcsponuzs to digoxin are illustrated in Figure I. There was a small but significant increase in the transcardiac fractional extraction of tritium-labeled norepinephrioc: however, there was no change in cardiac norepinephrine clcatancc (Table 2). In Gnnqr 8, there wcrc also no changes in arterial norcpinqrhrinc concertiration. total body norcpincphrine spillovcr or total body norepinephri~tr clearance after administration of digoxin. In contrast to tmdings in Group A. coronary sinus ncr-pinsphrine concctttration increased (fmm 20 i: ft.7 to 2.s f 0.7 nmnllita) after digoxin (p < 0.01) as did cardiac ntuegmepitrine spillover (from ICI Z 55 to 191 2 68 pmohmin)
(p < 0.05, Fig. 1). There was a trend to a decrease in cardiac norepinephrine clearance (p = 0.1, Table 2). In rhe rime conrrolgmup, there were no changes in catecholamine concentrations or spillover rates over a 30-min period (Table 2, Fig. 2). In the normal venMdarjiwtion group, there were also no catecholamine or norepinephrine spillover changes from control to 30 min after digoxin administration (Table 2, Fig. I).
Figure 1. Cardiac norepinephrine (NE) spillover response to digoxin in individual patients with group mean values (-) for 17 patients with congestivs heart failure (CHF) (Group A and Group 8) and 6 patients with normal ventricular function (Normal). Group A comprised 12 patients with severe failure and left ventricular end-diastolic pressure >14 mm Hg, and Group B, 5 patients with less severe failure and left ventricular end-diastolic pressure ~14 mm Hg. Spillover values are shown for the control measurement (C) and 30 min afler digoxin administration (0). “p c 0.0.5 and tp r, (I.01 versus the control value for within-group comparisons.
JACC Vol. 28. No. I July 1’?5Wl55-61
DIGOXIN
IN CONGWTlVE
NEWIUN ET AL. HEART FAILURE
159
+----
Ptgure 2. Cardiac norepinephrine (NE) spillover values in individual patients with group mean values t-) for the seven patients in the time control group. Spiker values are shows for the mntrol measurement (C) :tind at 30 min.
Behveen-group comparisons of cardiac norepioephrhe spfllover respnnses. The decrease in c;rtdiac norepinephrine spillover after dignxin administration in Group A was significantly different ttiim th. responses in the normal ventricular function group (p < 0.05) Group B (p < 0.01) and the time control group (p < 0.05). Group A and Group B differed in several hemodynamic variables (Table 1). A retrospectivefy defined analysis demonstrates that there appears to be- a continuous relation between the cardiac spillover response to digoxin and certain baseline variables. For example, linear regression analysis of all the patients with heart failure reveals a significant relation betsveen left ventricular ejection fraction and the change in cardiac norepinephrine spillover seen with dignxin (5.2x - l$8, r = 0.62, p = 0.01; Fig. 3). Similarly. there was a relation between cardiac index and the cardiac norepinephrine spillover response to digoxin, although in this case it was of borderline statistical significance (36x - 96, r = 0.41, p = 0.1). Dlhydroxypbenylglycal cardfac spillover. The cardiac spillover of tritiumdabeled DHPG tended to increase both in Group A (6,375 t 1,466 vs. 7,604 2 I.358 dpm/min, [control vs. digoxin, p = 0.211) and in the time control group (5,896 z 2,217 vs. $352 2 3,088 dput/min [control vs. 30 mitt, p = C.lg]). There were no changes in the cardiac spillover of endogenous DHPG in either group (data not shown). Digoxin levels. The serum digoxin level at 30 ruin was 4.g + 0.5 nmolihter in Group A, 4.3 5 0.7 nmohliter in Group Band 4.4 + 0.3 nmohliter in the normal ventricular function group.
Discussion The novel finding of the present investigation is that digoxin caused a red&on in cardiac norepinephrine spillover in patients with heart failure and decompensat ed hemodynamic status, as defined by elevated ventricular filling premres. Selecting patients based on elevated kft ventricular enddi+stolic presume resulted in a group with severe heart failure.
Ftgurs 3. Scattergram of change in cardiac norepktephrine (NE) spillover with iligoxiu versus left ventricular ejection fraction by radionuclide ventricutography, in patients with heart failure who received digoxin (Group A and Group B). Patients in Group A (elevated tilling pressures) had a tower cardiac index and higher pulmonary artery pressures than did patients in Group B (normaf filling pressures). Dioxin caused a cardiac sympathoiibitory effect only in Group A. We emphasize that this result was probably not based on ditherences in filling pressures alone. It appears liiely that the observed effect of digoxin in the patients in Group A resulted from the severity of their heart failure syndrome. Ilk concept is supported by the significant relation found between the sympathoinhibitory effect of digoxin and left ventricular ejection fraction. Because a decrease in sympathetic tone may have long-term beneficial effects in heart failure, our results may provide an explanation for earlier studies that demonstrated a favorable hemodynamic responx to digoxin during sustained therapy only in patients with alnormal hemodynamic status at baseline (30). keviuus invtstigatlons.The effect of cardiac gfycosides on peripheral sympathetic activity in patients with heart failure has been previously described. Although digoxin has been shown (16-19) to reduce plasma norepinephrine concentration, a general index of sympathetic tone, Goldsmith et al. (31) found no effect of a smafl noninotropic dose of cedilanid-D on total body norepinephrine spillover. We also found no change in total body norepinephrine spillover, suggesting that this meacure may be relatively insensitive to small changes in sympathetic outflow occurring in specific vascufar beds. Gn the other hand, Ferguson et al. (20) showed that de&no&k directfy decreased measured skeletal muscle sympathetic nerve act$ty. We considered it important to extend this observation to the heart because the symputhetiq nemous system is highly differentiated and effects observed in the periphery may not be representative of adrenergic drive to the heart. Consistent with the observation of Ferguson et al, in the periphery (20). we demonstrated that digoxin reduced sympathetic outflow to the heart in patients with severe congesthe heart failure. lWn!hl lcebrnisor l&s inktigation fciatsefi on ekrent sympathetic outflow to the heart. lk possible afferent
16G
NEW-t-ON ET AL. DIGOXIN IN CONGESTtVE
HEART FAILURE
mechanisms responsible for the observed dccrcasc in cardiac norepinephrinc spillover were not directly evaluated. Both the cent*ut and nerve terminal effects of glycosides cause increased noxcpinephrine rclcasc (32.33) and, therefore, do not explain our lindings. We speculate that the decrease in cardiac norepincphrine spillover was related to improved baroreceptor sensitivity. It is unlikely that it resulted from increased baroreceptrr stimulation, because there were no detectable inotropic or hemodynamic changes in response to digoxin. This finding suggests, but does not prove, that digoxin sensitized barorecaptor function. The reduction in cardiac norepinephrine spillover may have been limited to Group A because of a greater underlying impairment in ha.oreceptor control of sympathetic outflow m patients wit;) more advanced heart faihrre (34). Cardiac norepinephrine ,pillovcr increased in the heart failure group with normal hemodynamic smtus (Grout B). This unexpected result suggests that digoxin has the potential to increase cardiac sympathetic activity m patients with compensated heart failure. The present findings are only suggestive because the observation involves a small number of patieots nnd no control group with similar hemodynamic status was included. Furthermore, the mechanism of such an increase in cardiac norcpinephrine .spillover in patients with compensated heart failure is difficult to explain. Cardiac glycosides increase norcpincphrine release through both central sympathoexcitatory effects (32) and direct nerve terminal effects (33). These sympathoexcitatory effects have not been demonstrated in human studies and wern unlikely to have occurred in the present study because of the small dose of digoxin that was used. Ncvcrtheless, the possibility that digoxin increases cardiac norepinephrinc spillover in patients with compensated heart failure raises an important new hypothesis that merits further investigation. Cardiac norcpincphrinc spillovcr has been demonstrated in animal models to hc rcprcscntative of cardiac sympathetic nerve firing ralc (2h). Howcvsr. there are scvcral other variubles that may effect norepinephrine spillover, such as coronary blood Row, capi.lary permeability to norepinephrine (35) and norcpincphrine reuptake (25). In Group A, the group in which norepinephrine spillover was reduced, there was an increase in the fractional extraction of tritium-labeled norepinephrine consictent with increased rcuptake (36). However, the tritium-labeled DHPG spillover responxs in this group do nc’! ClIPr*tt an intrr:lsc in rcuptakc. Tritium-lahcled DHPG itriscs from the intracicuronal mct&olism of trilium-labeled norepincphrine that has been captured by neuronal reuptake (2S.26). Increased reuptake increases the intraneuronal concentration of tritium-lab&d norcpinephrine. causing an increase in the cardiac production of tritium-labeled DHKi. The tritium-labeled DHPG spillttver responses wcrc similar ir, Group A and the time controls group. suggc*ti:& that no augmenlation of Ieuptake occurred in Group A. The tir,s. dependent increa% chcrved in both groups h;r\ been P! own (3) to result from the intraneuronal accumulation of Mum-
JACC Vul. 3. No. I July l99fxlSS-61
labeled norcpinephrine with ongoing intravenous administration. Study limitations. Limitations to the experimental ap preach used in the present study must be considered. This was not a blinded or randomized study. However, all of the hemodynamic data were ana!yzed, and the patient assignments (Group A vs. Group B) were made before the biochemical analysis. Furthermore, the investigators performing the biochejlical analysis had no knowledge ‘of patient status. The absence of any change in norepinephrine spillover in the time control group supports our conclusion that the decrease in spillover in Group A resulted from the administration of digoxin and was not a function of time. We cannot rule out a time effect as the cause of the increase in cardiac norcpinephrine spillover observed in Group B. Our protocol did not include an assessment of baroreceptor function. However, afferent control of sympathetic outflow was considcrcd in the design of this experiment. A low dose of digoxin was chosen to minimize hemodynamic responses that would haie been likely to reduce sympathetic outflow without invoking any primary autonomic effects. In addition, a detailed and direct assessment of both left ven:ricular contractility and hemodyszmic variables was performed to define the stimulus to afferent control of sympathetic outflow, and to rule out any change in this stimulus with the infusion of digoxin. This study was an acute physiologic assessment of the cardiac sympathetic effects of a single dose of intravenous digoxin. The observed acute reduction in cardiac sympathetic activity does not provide information on whether similar effects would be present with long-term oral administration of digoxin. The recent report (19) of long-term benetitsof digoxin therapy on measures of heart rate variability in patients with heart failure suggests that the cardiac autonomic effects of digoxin therapy may persist with long-term use. Conclusions. A reduction in cardiac sympathetic activity was demonstrated after the admmistration of a small intravenous do= of Jig&n. This effect occurred in the absence cf any measurable motropic or hemodynamic changes, suggesting a primary autonomic mechanism. The finding of cardiac sympathoinhibition was limited to patients with severe heart failure, disturbed hemodynamir status and sympathetic activation, suggesting a potentially beneficial role for digoxin in modulating ncurohormonal activation in this group of patients.
JACC Vol. 28, No. I July 1996:155-61 Subpopuladons in nonfailing and Cailii human ve.nlrlc&r myocardium: mupIll of both recepor subtypea m muscle contraction and sefectivc Br-receplor down-regulation in heart failure. Cii Res 19&9%%3fN. 3. bin8111DL, Kent RL, Pamons 8, Conper G. Adrcnergic eUec!s on the biology of dw adult mammalian cardiueytc. Ciition 1992#799-804. 4. Meredith IT, Broughton A, Jennings GL, E&r MD. Evidence of a selecdi increase in cardiac sympathedc activity in patients with susmined ventricular arrhythmias. N EnSI J Med 1991;32%618-24. 5. Knye DM, Lelhovia 1, Jm GL, fkqia P, Bmqba A, E&r MD. Adverse mmquemes of high sympathetic nerwusaaivityinthecailing human hcari. J Am Coil Cardiol1995:5:26:1257-63. 4. Guest JA, GiBis RA. EBecl oCdi&ahs on carotid slnus lraroreceptor activity. Cii Res 1974;35:247-55. 7. Thames MD, Waickman LA, Abbuud FM. Senstimtion of cardii receptors (vagal atferents) by intracoronary acctylstrophanthidin. Am J Physiol l98B 239:H62S-35. S. Lopez JA. Timmis AD, Garan H. Homey Cl, Poweli WJ. ClLct of intracarobd administration of ouabain in doBs. Am J Physiol l9Sfl:W: H148-55. 9. Ferrari A, Gregorini f., Ferrari MC. Preti C, Mancia G. Dighalis and baroreceptor reflexes in man. Circumtion 19Sl;63279-85. 10. Schobel HP, Gren RIM, Roach PJ, Mark AL, Ferguson DW. Comrasting elfe& of digitabs and dobummine on baroreflea s~pathetic control in normal humans. Circulation 1991;84:1I M-29. I I. Ferguson DW, Ahboud FM, Mark AL. Selective impairement of barorcllcxmediated vasomnstrictor responses in patk nu with ventricular dysfunction. Cllaridliuu 19S4;69:451-60. 12. Watanabe AM. Di&alii and Ibc autonomic nervous system. J Am Coil Cardioll985;5:3SA-42A. 13. Gills RA. Cardiac sympathetic nerve activity: changes induced by ouabain and pmpranolol. Science 1969#6%-10. 14. Rebagay WR, Caldwell RW. Dilferenr patterns of autonomic nerve activity prodwed by a pohu vs. a neurrd cardiac &oside. J Pharmaml Esp Ther 1990;253:1ffB-4. IS. Sole MJ. Benedict CR Versteeg DHG, de Kloet ER. Digjtosin therapy partialfy resIores cardii camcholamine and brain serotonin metabolii in mngestive hem failure. J Mol Cell Cardiol1985;37:10%63. 16. Ribner HS. Phrcinski D4 Hsieh 4 et al. Acute effuts of diguin on total systemic vascular resistance in congestive hean failure due to dilated cardiomyopathy: a hemodynamic-hormonal study. Am J Cmdiil I9S5;W: 8%4X34. 17. Glwrghllde M, Hall V. La&r JB, Goldstein S. Conqarative hemodynamic and neurohormonal effec+r of intravenous capmpril and digosin and their combinations in paucms with severe heart failure. J Am Call Cardiol 198~lal.34-a. IfI. Van Veldhubn DJ, Man In’t Veld AJ, Dunselman PHJM, et al. Doubleblind plarebomntmlled smdy of ibopamine and digoain in p&ems with mdd lo moderate heart failure: resuhs of me Dulch lbopaorine Muldcenter Trial. J Am Coil Cardiol 1993;22:1564-73. 19. KNm H, Bigger IT. Goldsmith RI, Packer M. Effect of longterm digodn therapy on aulonomic function in patients with chronic hcan failure. J Am Coil Cardiol IW5.2.5:289-94.
DIGGXIN
fN CONGiLSWE
NBrnN BTAL HEART PAILURE
161
20. Feguson DW, Berg WI. Sanders JS. Roach PJ. Kempf IS, Kienrk MO. Sympathoinhibitory responses to ditalis glyaxidca in heart failure patients DIrea evidence from sympathetic neursl record@ Cbmdadoa I~~D$cI: 45-77. 21. bye DM Cambert GW, l.efkovim J. Morris M, Jemdngs G, Ealer MD. Nemochemical evidemx oC cardii sympametic activation and imremed central nemus system mxepinepbrine turnover in sewxe congestive heart failure. J Am Coll Cardiol 1994$Jz570-8. 22. Fenprson DW, Berg WJ, Sanders JS. Clinkal and hemodymunk correla&s of sympmhetic nerve activity in normal humans and patients with heart failvrc: evidencefmmdirectmicroneurographie~gs.JAmcdlcsntid IWO;lh:lm-34. 23. Leiibaeh WN. Wallin BG, Viia RG. Aylwnrd PC. SundloS 0. Mark AL CXrcct evidenm from imraneural rccordinfp for bmremed central ,syvnpa thetic outllow in patienu, with heart failure. Circuladon l9&73z913-9. 24. Esler M, Jennings G, Korner P, Blombery P, Sachar& N, Leonard P. Measurement of total end organ-specific norepinephrine kinetics in humans Am J Physiol1984;2/17:E21-8. 25. Meredith lT, Eiinbfer G, bmbm GW. Dewar EM. Jemriqrs GL. E&r MD. Cardiac sympathedc nervous activity in cxqesdve hcar~ Cailurc: evidence for increased neumnal nompinephrine release and preserved neuronal uptake, Circaladon lW3$lfl:l.36-41. 26. Ehenhofer G. Smoliih JJ, Cos HS. Eskr MD. Neuronal reuptake of norepineplninc snrf pmduction of dthydnnyphenylglycol by cardii sympa thetii nerves in the anaesthedzed oop Circulation lW1;84:13S4-63. 27. Esler M, Jackman G, Bob& A, et al. Determinanon of norepbmpbrine apparent r&me ram and clearance in humans. Life Sci 1979$+4&79 28. Mathey DG, Cbanejee K, Tyberg Iv, L&en I, Brundage B. Parmeky WW. Coronary sinm reflus: a umrce of error in tbe measurement of thermodilutinn coronmy sinus Row. Cirmdation 1971;57:778-86. 29. Ganz W, Tamma K, Minus HS. Dormso R, Suran HJC. Measurement of mmnary -6~ blood Bow by contimmus tbermodiluuon in man. Circulation 1971;&i%95. 30. Lee DC. Johnson RA, Bin&am 111 et al. Heart failure in outpadents: a tat&mid trial ofdignnin vemus placebo. N En@ J Med 19M:306699-705. 31. Goldsmith SR, Simon AB. Mi E. Effect of diytalis on norepincphnne kinetii in oqestive beart t%hrrc. J Am Cull Cardiol liM:L(tS58-63. 32. Garan K Smith TW. Powell WJ. The cemrsl nervous sptrm as a she of action for the coroaary vaurconstriior effect of digosm. J Clin Invest 197454: tw72. 33. Kranzhofer R Haass M, Kun T, Richard1 G. Scbomik A. Elfec~ of digimhs &cosides on numpincphrine release in me heart. Dual mechanism of aaim Circ Res 1Wl;h8:162&37. 34. Grima AE. Moe GW, Huwmd RI, Armsrmag PW. Recovery of aftenualed baroretlexscnsitivhy in conuiousdug after reversal of pacing induced heart failure. Cardiovw I& 19w;28%4-y0. 35. Co&neau D. Gore&y CA Bach GG, Rose CP. Effect of badrencrgic blokuk on m VW norcpincpbnnc rekzuc in canine heart. Am J Physiol 19ll4,2Jh:H?ll3Y? .36 Goldstein IX. Brush SE. Ei*. hofer G. kull R, Etr hf. In viva measuremcrn of ncuronal uptake of mucpmcphrinc in the human heart. f’oculalhm i’J%7XYJl-X.
155 HEART FAILURE
Digoxin Reduces Cardiac Sympathetic Actitity in Severe Congestive Heart Failure GARY E. NEWTON, MD, JEFFREY H. TONG, PHD,* ANNE M. SCHOFIELD, BScN, ANDREW D. BAINES, MD, PHD,* JOHN S. FLORAS, MD, DPHI~ FACC, JOHN D. PARKER, Toronto,
Ontario,
MD
Canaak
UAm call l2vllid 1996;2&155-61)
The syndrome of congestive heart failure is characterized by generalized and organ-specilk increases in efferent sympathetic activity, with the greatest increase in sympathetic out&w being directed to the heart (1). Increased sympathetic out&w to the heart may have several adverse effects in&xiii downregclation and desensitization of cxdiac beta-adrenergic receptors (2), impaired function and viability of cardiac myocytes (3) wd a predisposition to ventricular arrhythmias (4). Recent evidence (5) demonstrates that increased cardiac adrenergic drive is stron& predictive of mortality, at least in patients with heart failure undergoing evaluation for heart transplantation. Despite the adverse consequences of sympathetic activation
directed to the heart, the modulation of cardiac sympathetic outflow by medical therapy in the setting of heart failure remains largely unexplored. An agent that may favorably reduce central sympathetic outflow to the heart is digoxin. Cardiac glyadeq by increasing arterial baroreceptor and cardiac receptor nerve d&barge (6-ll), cause an &ease in efferent parasympathetic activity and a k&ction in central sympathetic outlbv (12). In animal models glycosides reduce cardiac effewnt sympathc’k activity (13,14) and increas cardii catecholamitse stores (15). In patients with heart failure digoxin reduces plasm norepkphrine concentration (1619), an approximation of gene&ized sympathetic tiivity. &diacg@&deshavealsokenshown(M)toreducesympatheticmxvetra&toskektalmusde,adkectmeasweof TyTtp&etic~tothe~.The’eilectofdigrednm !$ympdeticoutfknvloth!ehearlinhtmlanswithcoagestiwheart faihmremaiE3lmlmom. Predii which patients will have increased cardiac gympathetic o&low is dilkdt bemuse the afferent signals leading totbbdisaubanceate~~~defined(21).Lthesen~0f heart fake, efferent sympathetic activhy to skeletal musck is iSJUtWSiSlpulmonaryarterypressUrettSSd diredyrdatedto left vennicular lwlg pessures (2523). sid@. sympatbetk oll&vwtothehearias~bycafdiacnorepioephriae spjilaver,isalso p”ssures (2% 07354@97c96m5M m sa735-la97t%)fJtups
JAC’C VUI. 28. No. I July IYWxl.%-61 certain, cardiac kympalt.elic activalion in congeslivc ilcart failure is associated with elevated central filling prrssu,t .:. We hypothesized that digoxin would reduce cardiac sympathetic activity in patients with heart failure and underlying sympathetic activation. To test this hypothesis we studied a group of patients with heart failure and elevated filling pressures, because these patients have been reported (21-23) to have increased sympathetic activity. For comparison, we also, studied patients with heart failure and normal filling pressures and patients with normal ventricular function. To assess the effect of digoxin on sympathetic outflow IO the heart. we utilized the steady state radiotracer technique (24) to measure :he cardiac norcpinephrine spillover response to a single imr:!venous dose of digoxin.
Methods Patients. Twenty-four patients with heart failure (22 men, 2 women; mean age 2: SEM 58 f 2 ycan) with a left ventricular ejection fruction ~35% (radionuclide angiography [n = lb], cchocardiography [n = I]) participated in this study. Four patients were in New York &art Association functional class II and 21) were in class 111 c: IV. Twelve patients with elcvatcd left ventricular end-diastolic pressure (>14 mm Hg) were designated as Grrncp A. The etiolog of heart failure 1~1this group was ischrnric in nine and idiopathic in three patients. TrGIImenl included diuretic drugs (II = Id), angiotensclconverting enqme inhibitors (n 7 IO), digoxin (n = 9) and amiodarorre (n := 2). Five palicnts with normal filling pressuro (left ventricular end-diastolic pressure 514 mm Hg, Group R) also received digoxin. The etiology of heart failure in this group was ischl:mic in thlce and itliclpathir in two patients. ‘frcatmcrtf included diuretic drugs (n L ?). angiotcnsinconvelting enzyme inbihitors (n 2 2) ilnd digoxin (I: = 2). Scvcn patients with hem!Klynamic dat:B similar to those of Group B (Gnru co~rr~!gro~p) underwent an identical protocol without the administration of digoxin. The etiology of heart failure in this group was ischemic in three and idiopathic in four patients. Treatment included diuretic drugs (n .= 7). angiotensinconvcrting cnaymc inhihiton (n 2 h), digoxin (II .7) and amitdurcmc (n 2). Paticnls participating in thr dig&n protocol who were rccciving maintenance digox& thcrapj had this medic&ion withheld for a minimrlm of 7 days before the sludy date. The nor& r’c~nr~~~~br~~nc~/ic. 1grvrrp included five men and ow woman (mu. I LI~C 50 ! 5 yciln). Three paticms had ct)t~on;l~ artcry JIwa’iL’. bul M)IIC hild a hishfry of myncardial mfarstion. Medicid thrrapy included hta-adrcncrgic blocking agents (II = 2). calcium channel hlockmg agents (n = 2). and an angiotensinccmvertting enzyme inhihitor (n .= !), IX protocol was approved by rhe phical Review Committed for Human Ewiin;intation of Ihc :‘nivcrsicj of TG romo. Writiea informed conwnt y/a\, trhtaincd in ,111cuses“~. Shady pr&cai. Diagntatic right and fcft hc:lri *.atheterilation, without udation. was performed fnr.1 the femoral ap proaeh. A 7F ?lumnlil~~~~m caarlctcr (rvp‘ (‘G-?tJ-Wf3.
Wchstcr Ishoratories) was then placed in the coronary sinus from an antecubital vein for dew measurements and sampling. A 7F micromanometer-tipped catheter (Milbu Industries) was placed in the left ventricle for assessment of contractility (peak positive first derivative of left ventricular pressure (dP/dt) and filling pressures. The patient was then left undisturbed for a minimum of 20 min and until an intravenous infusion of tritium-labeled norepinephrine reached steady state concentration in plasma. Al that point, hemodynamic variables and total body and cardiac norepinephrine spillover rates were assessed. Digoxin (0.25 mg in IO ml 01 5% dextrose in water), was then administered intravenously over 6 min with use of a Harvard pump. Hemodynamic and norepill. ihrine spillover mcasuremcnts were repeated 30 mm later. In the time control group, the same hemodynamic and spillover i;*sasurements were performed at baseline and repeated at 30 min. Patients were classified by left ventricular enddiastolic pressure measurements into either Group A (>I4 mm Hg) or Group B (-14 mm Hg) on the basis of hemodynamic data recorded before administration of digoxin. This catego:ization was performed before the biochemical analysis. Norepinephrine spillover merwwemeots. Sympathetic activity was estimated according to the radiotracer technique of Esier et al. (24). This method provides the rate at which norepincphrine appears in plasma and is therefore an indirect measure of effcrcnt sympathetic nervous activity. To assess norepincphrine spillover, a ...~~dy state plasma concentration of :I tracer dose of tritium-labeled norepincphrine is required. Total hrldy norepincphrine spillovcr is cala:latcd as follows (L24): Tot:11 h)dy NE ‘:pillwer (pmol/min) [‘IlINE infusion rate (dpm/min) = Plasma NE &cific activity (dpmipmol) where NE = norcpinephrine, [JH]NE epincphrine, and dpm = disintegrations labclrd norcpincphrine. C’ardiac norcpincphrinl: spilltwcr is the Fick principle, with a correction tritium-labeled norepinephrine across
’
= tritiilm-labeled nor. per minute of tritiumcalculated according for the extraction the heart (1,24,25):
IO of
whcrc Nt.,.,, - tramcardiac fractional extraction of tritiumlab&d norrpinephrinc. (JH]NE,,,, and t’HINE,, = arterial and coronary sinus tritium-labeled norepinephrine conccnrralions. respectively. NE, and NE,, = coronary sinus and -rtcr~al norcpincphrinc concentr&ons. respcctivcly, ,UI~ I’SPI: = coronary sinus p&ma How calcufatcd from the hr‘nrattrrit and coronary sinus blood Row. A nwc dctaikd cbaluation t$ cardiac mwspinephriw kineticswus pwf&xrnwd in Md:hc I2 patients in Gruup A sld
JACC Vol. 24 No. 1 July 1996:155;61
in all 7 in the time control group. In these patients, the cardiac spillover of the intraneuronal norepinephrine metabnlitc, dihydroxyphenylglycol (DHPG), was calculated as the product of its venoarteriaf gradient and coronary sinus blood flow (25.26). Cardiac spillover of tritium-labeled DHPG was also calculated by the same method (25.26). Tritium-labeled DHPG spillover, which can only occur aa a result of neuronal uptake of tritium-labeled nore,inephrim, provides an index of neuronal re-uptake of norepi’ ephrine (25,26). Radiotmcer pw~watioo and ioliis!sn. Trifium-iabekd norepinephrine (L-I?J,~‘H]NE, New England Nuclear) was diluted in sterile 0.2~moffliter acetic acid in saline solution with IO-mmoi/liter ascorbate, sterilized by filtration through a 0.22~pm Millex-GV filter and packaged in 4-ml aliquots that were stored at -70°C until use. The radiotracer was administered intravenously as a IQ@Ji priming bolus followed by a continuous infusion of 1 Lo 1.2 PCiimin. Control mcdsurements were performed ~60 min after initiation of the radiotracer infusion to ensure steady state concentration in plasma (27). Analysis of plasma cateehulamines. Samples were transferred immediately to ice-chilled tubes containing an anticoagulant (ethylenediaminetetraacetic acid). The plasma was separated by centrifugation at 4°C and then stored at -70°C until assayed. After the addition of dihydroxybenzlamine (DHBA) as an internal standard, the catechols were extracted by adsorption on alumina. After elution from the alumina by the addition of O.Zmol/liter perchloric acid, the cacechols were separated by isocratic ion pair chromatography on a reversed phase column. The high performance liquid chromatography system consisted of a multisolvent delivery system model 6OOE pump, a model 7171 automatic sample injector (Waters Associates) and a Zorbax %-Cl8 4.6 X 1%mm column (Chromatographic Specialties Inc., Brockville, Ontario. Canada). Detection of the compounds was performed by a model Coulochem II
DIOOXIN
IN CONGESTIVE
Table I. Baseline Characteristics
NbWiDN ET AL HEART FAILURE
157
of the 30 Patients Patients With Heart Failure
Patients With Normal Ventricular Fmmion (n = 6) .-.-.LVEF sv c 5 HR 6v.!3 14: 2 PAmcm Cl 2.7 t. 0.1 L dilJP II r2 153 2 10 F-h FA PP 7725 i377tw tdP/dt CANFSP III3 f 41
Group A (n = 12) ?I + 1. tl6.tJ 30 t 3’ l.x+u.I* 25 t 2’ Es+7 56 t 4’ vwtvw 263 ? 70
GwB (n = 5) 31 + ,(4 : II 2 2.7 + Vtl 137: 76 + l020? 151 t
3’ 4 2t 0.3t 7 4 122 5s
Time ContFol Gmup (n = 7) 152 I’ sn? IO 33 + 5’ 1.8 ? 0.2 24 t 4” 1t1+7 5427 706 ‘i 54’ 33v ? 73
‘p < 0.0s vcmu nwmal gr”“uup.tp . UM YenIl\ alup A. Left ventrictdw ettd-dimtalic pwsure (LS’EDP [mm Hgl) was not annpted beween Group A pqd Croup 6. Data presented are mean value t SEM. CANESP = cardiac norepmephrine spillover @moUrnin); Cl F cardiac index (litem/min per m*): FA PP = femural artery pulse pressure (mm Hgj: F& = femoral artmy systobc prwure (mm Hg); HR = bean rate (beztsimin); LVEF = left vetttktdat ejection fraction (a); P&.. = mean pulmonary attery pressure (mm fig); tdP/dt = peak positive lint derivative of left ventricular pressure (mm Hgk).
tion) were recorded on a strip chart recorder. Measurements were obtained by averaging 2 I5 cardiac cycles. The coronary sinus thermodilution flow catheter was positioned under lluoroswpic guidance and correct placement was confirmed with radiographic cOntrast injection. The external thermistor was advanced 22 cm past the ostium of the coronary sinus to avoid reflux of right atriil blood (28). A amstant position was maintained during the study by comparison with anatomic landmarks. Room temperature 5% dextrose in water was infused at 35 ml/min by Harvard pump for coronary blood flow measurements, which were performed in triplicate according to the method of Ganz et al. (29). Statlstkal natysi~. All data are presented as mean value -t SEM. Statistical analysis was performed in SAS (SAS release 6.10, SAS Institute Inc.). Between-group comparisons of baseline characteristics were ma& by a one-way analysis of variance adjusting for multiple comparisons (Bonferroni). Within-group and between-group comparisons of changes between control measurements and 30 min after digoxin administration (and time control) measurements were made using a multivariate anatysii of variamc (MANOVA) with appropriate contrast statements for all mmparisons. A p value of less than O.O!i was required for statistical signitkance.
l
Results
Baselii characteristics (Table I). The 12 patients with heart failure in Group A (left ventricular end-diastolic pressure ; I4 nn Hg) had a mean left ventricular enddiastolic pressure of 25 C 2 mm Hg (rzngc I6 to 3x). The tive patiena in, Grarpr3(cfiwXtricu!ai&-diast&peswe~I4mmHg)had ameankftventrkularer&%astolicpressumof9~ ImmHg (range4tol2).InamparisonwitbGroupB.thepatientsin
158
NEWTON ET Al.. DlODXlN IN CONGESTIVE
JACC Vol. 28, No. I Jul) 1%155-61
HEART FAILURti
Table 2, Hemodynamic, Plasma (Btecholamine b---e-
and Norepinephrine
Spitlower Responses
Pv~irnlr Wilh Normal Vcntriculur Fun&n (n b) ---c Dig
Gnq A (n 12) ------
--
Pntirntr With llcarr Fadurc
-~ c
Dig
---_-----c.
:inmp B In -. S)
Time Ctihlml Group (w= 7) Dig
c
30 min
HR 69 -r 3 69 % 4 x6+5 MSi 5 8454 Xl 24 98 2 10 % 2 IO LVEDP II +-2 102 1 2522 25-3 9+1 72 I 24 2 4 23 5 4 h‘% 153 _c IO 155 + 8 12827 132 5 6 137 + 7 141 2 IO 11827 118~h FA PP 77 -r 5 80%6 56 + 4 59 _+4 7b “- 4 79i-9 54 c 7 56 f 7 +dPidl 3.377 +_99 1381 2 92 99Glc9u 1027 + 55 1020 + 121 1053 f 131 7Ob t 54 716 c 58 CSBF I43 f II 14x L 15 176 f !I IbS .Z ?I 17s t I4 Ihl + 14 185 2 35 IYI ?z34 NL,, 1.1 :o..l I,/ ? 0.2 2.2 5 0.4 I.0 i 0.3 I .4 ! 0.4 I .4 t (I.4 3.2 ? 0.X 3.0 c 0.x W I.4 t 0.4 1.4 t Il.3 3b? 1.1 3.2 r 1.w 2.0 c (I.7 2.5 f 0.W 4.5 ? o.ii t.b ?: 0.8 TBNESP 3.23 t Oh9 3.34 t 0.5v 4.02 ‘f 0.68 4 44 f Rb5 4Ih? 1.m 4.62 * I.32 b.25 ? 0.57 6.44 z 0.67 NE,., (I.71 i: (1.07 0.7l1 t o.oh 031 1 0.03 0.61 ? rl.tnt o.sv * 0.05 03x 3 0.07 0.54 f: 0.07 0.50 + 0.07 CANECL 71 t I4 7H t II ho t x 61 + 4 bV f 7 61 *_ 7 56 i- 5 5b+ 5 CANESP 103 .r 41 vu c 37 2h3 ? 70 2IX + h3 I51 + 55 IV1 t 6X' 339 c 73 339 2 65 -..-*p < OJI5 and tp < (I,01 wrsus clmtwl value for xithiwgroup compwium,. Data presented we mean value + SEM. C = conlml; CANECL = cardiac norepincphrinc clearance (ml/mm): CSBF = coronary 6iw.s hltd Row (mllmin); IJig = digoxin; NE,, and Nf&,. = arterial and coronary rinus norepinephrine waeeatration (nmcl4iter. respeclivrly): NE,,, = tramcardiac vu dm-labeled norcpinephrine cxtrdclion ($6): TBNESP = lotal body norepinephrinc qpillover (nmol/min): other ahhrtvtatmn~ 4s m Talk 1.
Group A had significantly lower values for cardiac index and higher mean pulmonary artery pressures. Although the mean value for cardiac norepinephrine spillover was higher in Group A than in Group B, the difference was not significant. Baseline hemodynamic and nemochemical values in the time control group with heart failure were similar to those in Group A. Baseline characteristics for all groups are summarized in Table 1. Humudynamic rcspunses (Tabie 2). There were no changes in any hemodynamic variable after administration of digoxin in Group A, Group Band the normal ventricular function group. There were also no hcmodynamic changes over 30 min in the titnc control group. Crrdinc and tutal body norepinephrlne spillover responses (Table 2. Fig. I and 2). In G”tuup A, there were no significant changes in arterial uorepinephrine concentration, total body norepinephrine spillover or total body norepinephrine clearance after administration of digoxin. However, coronary sinus norepinephrine concentration decreased from 3.6 t 1.1 to 3.2 4 1.0 nmol/iiter after digoxin (p < 0.01). Similarly. cardiac norepinephrine spillover dccreascd from 263 i’ 70 to 21X L 62 pmol/min after digoxin (p e.: Wttl). The indivtduat cardiac norcpincphrine spillover rcsponuzs to digoxin are illustrated in Figure I. There was a small but significant increase in the transcardiac fractional extraction of tritium-labeled norepinephrioc: however, there was no change in cardiac norepinephrine clcatancc (Table 2). In Gnnqr 8, there wcrc also no changes in arterial norcpinqrhrinc concertiration. total body norcpincphrine spillovcr or total body norepinephri~tr clearance after administration of digoxin. In contrast to tmdings in Group A. coronary sinus ncr-pinsphrine concctttration increased (fmm 20 i: ft.7 to 2.s f 0.7 nmnllita) after digoxin (p < 0.01) as did cardiac ntuegmepitrine spillover (from ICI Z 55 to 191 2 68 pmohmin)
(p < 0.05, Fig. 1). There was a trend to a decrease in cardiac norepinephrine clearance (p = 0.1, Table 2). In rhe rime conrrolgmup, there were no changes in catecholamine concentrations or spillover rates over a 30-min period (Table 2, Fig. 2). In the normal venMdarjiwtion group, there were also no catecholamine or norepinephrine spillover changes from control to 30 min after digoxin administration (Table 2, Fig. I).
Figure 1. Cardiac norepinephrine (NE) spillover response to digoxin in individual patients with group mean values (-) for 17 patients with congestivs heart failure (CHF) (Group A and Group 8) and 6 patients with normal ventricular function (Normal). Group A comprised 12 patients with severe failure and left ventricular end-diastolic pressure >14 mm Hg, and Group B, 5 patients with less severe failure and left ventricular end-diastolic pressure ~14 mm Hg. Spillover values are shown for the control measurement (C) and 30 min afler digoxin administration (0). “p c 0.0.5 and tp r, (I.01 versus the control value for within-group comparisons.
JACC Vol. 28. No. I July 1’?5Wl55-61
DIGOXIN
IN CONGWTlVE
NEWIUN ET AL. HEART FAILURE
159
+----
Ptgure 2. Cardiac norepinephrine (NE) spillover values in individual patients with group mean values t-) for the seven patients in the time control group. Spiker values are shows for the mntrol measurement (C) :tind at 30 min.
Behveen-group comparisons of cardiac norepioephrhe spfllover respnnses. The decrease in c;rtdiac norepinephrine spillover after dignxin administration in Group A was significantly different ttiim th. responses in the normal ventricular function group (p < 0.05) Group B (p < 0.01) and the time control group (p < 0.05). Group A and Group B differed in several hemodynamic variables (Table 1). A retrospectivefy defined analysis demonstrates that there appears to be- a continuous relation between the cardiac spillover response to digoxin and certain baseline variables. For example, linear regression analysis of all the patients with heart failure reveals a significant relation betsveen left ventricular ejection fraction and the change in cardiac norepinephrine spillover seen with dignxin (5.2x - l$8, r = 0.62, p = 0.01; Fig. 3). Similarly. there was a relation between cardiac index and the cardiac norepinephrine spillover response to digoxin, although in this case it was of borderline statistical significance (36x - 96, r = 0.41, p = 0.1). Dlhydroxypbenylglycal cardfac spillover. The cardiac spillover of tritiumdabeled DHPG tended to increase both in Group A (6,375 t 1,466 vs. 7,604 2 I.358 dpm/min, [control vs. digoxin, p = 0.211) and in the time control group (5,896 z 2,217 vs. $352 2 3,088 dput/min [control vs. 30 mitt, p = C.lg]). There were no changes in the cardiac spillover of endogenous DHPG in either group (data not shown). Digoxin levels. The serum digoxin level at 30 ruin was 4.g + 0.5 nmolihter in Group A, 4.3 5 0.7 nmohliter in Group Band 4.4 + 0.3 nmohliter in the normal ventricular function group.
Discussion The novel finding of the present investigation is that digoxin caused a red&on in cardiac norepinephrine spillover in patients with heart failure and decompensat ed hemodynamic status, as defined by elevated ventricular filling premres. Selecting patients based on elevated kft ventricular enddi+stolic presume resulted in a group with severe heart failure.
Ftgurs 3. Scattergram of change in cardiac norepktephrine (NE) spillover with iligoxiu versus left ventricular ejection fraction by radionuclide ventricutography, in patients with heart failure who received digoxin (Group A and Group B). Patients in Group A (elevated tilling pressures) had a tower cardiac index and higher pulmonary artery pressures than did patients in Group B (normaf filling pressures). Dioxin caused a cardiac sympathoiibitory effect only in Group A. We emphasize that this result was probably not based on ditherences in filling pressures alone. It appears liiely that the observed effect of digoxin in the patients in Group A resulted from the severity of their heart failure syndrome. Ilk concept is supported by the significant relation found between the sympathoinhibitory effect of digoxin and left ventricular ejection fraction. Because a decrease in sympathetic tone may have long-term beneficial effects in heart failure, our results may provide an explanation for earlier studies that demonstrated a favorable hemodynamic responx to digoxin during sustained therapy only in patients with alnormal hemodynamic status at baseline (30). keviuus invtstigatlons.The effect of cardiac gfycosides on peripheral sympathetic activity in patients with heart failure has been previously described. Although digoxin has been shown (16-19) to reduce plasma norepinephrine concentration, a general index of sympathetic tone, Goldsmith et al. (31) found no effect of a smafl noninotropic dose of cedilanid-D on total body norepinephrine spillover. We also found no change in total body norepinephrine spillover, suggesting that this meacure may be relatively insensitive to small changes in sympathetic outflow occurring in specific vascufar beds. Gn the other hand, Ferguson et al. (20) showed that de&no&k directfy decreased measured skeletal muscle sympathetic nerve act$ty. We considered it important to extend this observation to the heart because the symputhetiq nemous system is highly differentiated and effects observed in the periphery may not be representative of adrenergic drive to the heart. Consistent with the observation of Ferguson et al, in the periphery (20). we demonstrated that digoxin reduced sympathetic outflow to the heart in patients with severe congesthe heart failure. lWn!hl lcebrnisor l&s inktigation fciatsefi on ekrent sympathetic outflow to the heart. lk possible afferent
16G
NEW-t-ON ET AL. DIGOXIN IN CONGESTtVE
HEART FAILURE
mechanisms responsible for the observed dccrcasc in cardiac norepinephrinc spillover were not directly evaluated. Both the cent*ut and nerve terminal effects of glycosides cause increased noxcpinephrine rclcasc (32.33) and, therefore, do not explain our lindings. We speculate that the decrease in cardiac norepincphrine spillover was related to improved baroreceptor sensitivity. It is unlikely that it resulted from increased baroreceptrr stimulation, because there were no detectable inotropic or hemodynamic changes in response to digoxin. This finding suggests, but does not prove, that digoxin sensitized barorecaptor function. The reduction in cardiac norepinephrine spillover may have been limited to Group A because of a greater underlying impairment in ha.oreceptor control of sympathetic outflow m patients wit;) more advanced heart faihrre (34). Cardiac norepinephrine ,pillovcr increased in the heart failure group with normal hemodynamic smtus (Grout B). This unexpected result suggests that digoxin has the potential to increase cardiac sympathetic activity m patients with compensated heart failure. The present findings are only suggestive because the observation involves a small number of patieots nnd no control group with similar hemodynamic status was included. Furthermore, the mechanism of such an increase in cardiac norcpinephrine .spillover in patients with compensated heart failure is difficult to explain. Cardiac glycosides increase norcpincphrine release through both central sympathoexcitatory effects (32) and direct nerve terminal effects (33). These sympathoexcitatory effects have not been demonstrated in human studies and wern unlikely to have occurred in the present study because of the small dose of digoxin that was used. Ncvcrtheless, the possibility that digoxin increases cardiac norepinephrinc spillover in patients with compensated heart failure raises an important new hypothesis that merits further investigation. Cardiac norcpincphrinc spillovcr has been demonstrated in animal models to hc rcprcscntative of cardiac sympathetic nerve firing ralc (2h). Howcvsr. there are scvcral other variubles that may effect norepinephrine spillover, such as coronary blood Row, capi.lary permeability to norepinephrine (35) and norcpincphrine reuptake (25). In Group A, the group in which norepinephrine spillover was reduced, there was an increase in the fractional extraction of tritium-labeled norepinephrine consictent with increased rcuptake (36). However, the tritium-labeled DHPG spillover responxs in this group do nc’! ClIPr*tt an intrr:lsc in rcuptakc. Tritium-lahcled DHPG itriscs from the intracicuronal mct&olism of trilium-labeled norepincphrine that has been captured by neuronal reuptake (2S.26). Increased reuptake increases the intraneuronal concentration of tritium-lab&d norcpinephrine. causing an increase in the cardiac production of tritium-labeled DHKi. The tritium-labeled DHPG spillttver responses wcrc similar ir, Group A and the time controls group. suggc*ti:& that no augmenlation of Ieuptake occurred in Group A. The tir,s. dependent increa% chcrved in both groups h;r\ been P! own (3) to result from the intraneuronal accumulation of Mum-
JACC Vul. 3. No. I July l99fxlSS-61
labeled norcpinephrine with ongoing intravenous administration. Study limitations. Limitations to the experimental ap preach used in the present study must be considered. This was not a blinded or randomized study. However, all of the hemodynamic data were ana!yzed, and the patient assignments (Group A vs. Group B) were made before the biochemical analysis. Furthermore, the investigators performing the biochejlical analysis had no knowledge ‘of patient status. The absence of any change in norepinephrine spillover in the time control group supports our conclusion that the decrease in spillover in Group A resulted from the administration of digoxin and was not a function of time. We cannot rule out a time effect as the cause of the increase in cardiac norcpinephrine spillover observed in Group B. Our protocol did not include an assessment of baroreceptor function. However, afferent control of sympathetic outflow was considcrcd in the design of this experiment. A low dose of digoxin was chosen to minimize hemodynamic responses that would haie been likely to reduce sympathetic outflow without invoking any primary autonomic effects. In addition, a detailed and direct assessment of both left ven:ricular contractility and hemodyszmic variables was performed to define the stimulus to afferent control of sympathetic outflow, and to rule out any change in this stimulus with the infusion of digoxin. This study was an acute physiologic assessment of the cardiac sympathetic effects of a single dose of intravenous digoxin. The observed acute reduction in cardiac sympathetic activity does not provide information on whether similar effects would be present with long-term oral administration of digoxin. The recent report (19) of long-term benetitsof digoxin therapy on measures of heart rate variability in patients with heart failure suggests that the cardiac autonomic effects of digoxin therapy may persist with long-term use. Conclusions. A reduction in cardiac sympathetic activity was demonstrated after the admmistration of a small intravenous do= of Jig&n. This effect occurred in the absence cf any measurable motropic or hemodynamic changes, suggesting a primary autonomic mechanism. The finding of cardiac sympathoinhibition was limited to patients with severe heart failure, disturbed hemodynamir status and sympathetic activation, suggesting a potentially beneficial role for digoxin in modulating ncurohormonal activation in this group of patients.
JACC Vol. 28, No. I July 1996:155-61 Subpopuladons in nonfailing and Cailii human ve.nlrlc&r myocardium: mupIll of both recepor subtypea m muscle contraction and sefectivc Br-receplor down-regulation in heart failure. Cii Res 19&9%%3fN. 3. bin8111DL, Kent RL, Pamons 8, Conper G. Adrcnergic eUec!s on the biology of dw adult mammalian cardiueytc. Ciition 1992#799-804. 4. Meredith IT, Broughton A, Jennings GL, E&r MD. Evidence of a selecdi increase in cardiac sympathedc activity in patients with susmined ventricular arrhythmias. N EnSI J Med 1991;32%618-24. 5. Knye DM, Lelhovia 1, Jm GL, fkqia P, Bmqba A, E&r MD. Adverse mmquemes of high sympathetic nerwusaaivityinthecailing human hcari. J Am Coil Cardiol1995:5:26:1257-63. 4. Guest JA, GiBis RA. EBecl oCdi&ahs on carotid slnus lraroreceptor activity. Cii Res 1974;35:247-55. 7. Thames MD, Waickman LA, Abbuud FM. Senstimtion of cardii receptors (vagal atferents) by intracoronary acctylstrophanthidin. Am J Physiol l98B 239:H62S-35. S. Lopez JA. Timmis AD, Garan H. Homey Cl, Poweli WJ. ClLct of intracarobd administration of ouabain in doBs. Am J Physiol l9Sfl:W: H148-55. 9. Ferrari A, Gregorini f., Ferrari MC. Preti C, Mancia G. Dighalis and baroreceptor reflexes in man. Circumtion 19Sl;63279-85. 10. Schobel HP, Gren RIM, Roach PJ, Mark AL, Ferguson DW. Comrasting elfe& of digitabs and dobummine on baroreflea s~pathetic control in normal humans. Circulation 1991;84:1I M-29. I I. Ferguson DW, Ahboud FM, Mark AL. Selective impairement of barorcllcxmediated vasomnstrictor responses in patk nu with ventricular dysfunction. Cllaridliuu 19S4;69:451-60. 12. Watanabe AM. Di&alii and Ibc autonomic nervous system. J Am Coil Cardioll985;5:3SA-42A. 13. Gills RA. Cardiac sympathetic nerve activity: changes induced by ouabain and pmpranolol. Science 1969#6%-10. 14. Rebagay WR, Caldwell RW. Dilferenr patterns of autonomic nerve activity prodwed by a pohu vs. a neurrd cardiac &oside. J Pharmaml Esp Ther 1990;253:1ffB-4. IS. Sole MJ. Benedict CR Versteeg DHG, de Kloet ER. Digjtosin therapy partialfy resIores cardii camcholamine and brain serotonin metabolii in mngestive hem failure. J Mol Cell Cardiol1985;37:10%63. 16. Ribner HS. Phrcinski D4 Hsieh 4 et al. Acute effuts of diguin on total systemic vascular resistance in congestive hean failure due to dilated cardiomyopathy: a hemodynamic-hormonal study. Am J Cmdiil I9S5;W: 8%4X34. 17. Glwrghllde M, Hall V. La&r JB, Goldstein S. Conqarative hemodynamic and neurohormonal effec+r of intravenous capmpril and digosin and their combinations in paucms with severe heart failure. J Am Call Cardiol 198~lal.34-a. IfI. Van Veldhubn DJ, Man In’t Veld AJ, Dunselman PHJM, et al. Doubleblind plarebomntmlled smdy of ibopamine and digoain in p&ems with mdd lo moderate heart failure: resuhs of me Dulch lbopaorine Muldcenter Trial. J Am Coil Cardiol 1993;22:1564-73. 19. KNm H, Bigger IT. Goldsmith RI, Packer M. Effect of longterm digodn therapy on aulonomic function in patients with chronic hcan failure. J Am Coil Cardiol IW5.2.5:289-94.
DIGGXIN
fN CONGiLSWE
NBrnN BTAL HEART PAILURE
161
20. Feguson DW, Berg WI. Sanders JS. Roach PJ. Kempf IS, Kienrk MO. Sympathoinhibitory responses to ditalis glyaxidca in heart failure patients DIrea evidence from sympathetic neursl record@ Cbmdadoa I~~D$cI: 45-77. 21. bye DM Cambert GW, l.efkovim J. Morris M, Jemdngs G, Ealer MD. Nemochemical evidemx oC cardii sympametic activation and imremed central nemus system mxepinepbrine turnover in sewxe congestive heart failure. J Am Coll Cardiol 1994$Jz570-8. 22. Fenprson DW, Berg WJ, Sanders JS. Clinkal and hemodymunk correla&s of sympmhetic nerve activity in normal humans and patients with heart failvrc: evidencefmmdirectmicroneurographie~gs.JAmcdlcsntid IWO;lh:lm-34. 23. Leiibaeh WN. Wallin BG, Viia RG. Aylwnrd PC. SundloS 0. Mark AL CXrcct evidenm from imraneural rccordinfp for bmremed central ,syvnpa thetic outllow in patienu, with heart failure. Circuladon l9&73z913-9. 24. Esler M, Jennings G, Korner P, Blombery P, Sachar& N, Leonard P. Measurement of total end organ-specific norepinephrine kinetics in humans Am J Physiol1984;2/17:E21-8. 25. Meredith lT, Eiinbfer G, bmbm GW. Dewar EM. Jemriqrs GL. E&r MD. Cardiac sympathedc nervous activity in cxqesdve hcar~ Cailurc: evidence for increased neumnal nompinephrine release and preserved neuronal uptake, Circaladon lW3$lfl:l.36-41. 26. Ehenhofer G. Smoliih JJ, Cos HS. Eskr MD. Neuronal reuptake of norepineplninc snrf pmduction of dthydnnyphenylglycol by cardii sympa thetii nerves in the anaesthedzed oop Circulation lW1;84:13S4-63. 27. Esler M, Jackman G, Bob& A, et al. Determinanon of norepbmpbrine apparent r&me ram and clearance in humans. Life Sci 1979$+4&79 28. Mathey DG, Cbanejee K, Tyberg Iv, L&en I, Brundage B. Parmeky WW. Coronary sinm reflus: a umrce of error in tbe measurement of thermodilutinn coronmy sinus Row. Cirmdation 1971;57:778-86. 29. Ganz W, Tamma K, Minus HS. Dormso R, Suran HJC. Measurement of mmnary -6~ blood Bow by contimmus tbermodiluuon in man. Circulation 1971;&i%95. 30. Lee DC. Johnson RA, Bin&am 111 et al. Heart failure in outpadents: a tat&mid trial ofdignnin vemus placebo. N En@ J Med 19M:306699-705. 31. Goldsmith SR, Simon AB. Mi E. Effect of diytalis on norepincphnne kinetii in oqestive beart t%hrrc. J Am Cull Cardiol liM:L(tS58-63. 32. Garan K Smith TW. Powell WJ. The cemrsl nervous sptrm as a she of action for the coroaary vaurconstriior effect of digosm. J Clin Invest 197454: tw72. 33. Kranzhofer R Haass M, Kun T, Richard1 G. Scbomik A. Elfec~ of digimhs &cosides on numpincphrine release in me heart. Dual mechanism of aaim Circ Res 1Wl;h8:162&37. 34. Grima AE. Moe GW, Huwmd RI, Armsrmag PW. Recovery of aftenualed baroretlexscnsitivhy in conuiousdug after reversal of pacing induced heart failure. Cardiovw I& 19w;28%4-y0. 35. Co&neau D. Gore&y CA Bach GG, Rose CP. Effect of badrencrgic blokuk on m VW norcpincpbnnc rekzuc in canine heart. Am J Physiol 19ll4,2Jh:H?ll3Y? .36 Goldstein IX. Brush SE. Ei*. hofer G. kull R, Etr hf. In viva measuremcrn of ncuronal uptake of mucpmcphrinc in the human heart. f’oculalhm i’J%7XYJl-X.